AI Day provided an opportunity for students, faculty and community members to come together and discuss the latest innovations in artificial intelligence.
Artificial intelligence (AI) is becoming increasingly engrained in industries, research, education and more, making discussions of ethicality and practicality important as technologies advance at a rapid pace. Such were the conditions that created the first-ever AI Day at Brock.
Brock’s Computer Science Department held the event on Nov. 9, which promised “to be a confluence of brilliant minds, groundbreaking ideas and the latest advancements in the realm of AI.”
The day began with some opening remarks and a brief presentation on AI initiatives at Brock, including the proposed Centre of Artificial Intelligence. Following this, Dr. Jad Kabbara gave a keynote speech on human-centred AI and related projects at MIT, where he conducts research for the Center of Constructive Communication.
Following his speech, much of the day entailed technical presentations on recent advancements in AI in realms such as public transit, drug design and gas wells.
Research posters were displayed on the upper floor of the conference room, many of which discussed the various usages of Large Language Models (LLMs), the most prominent element of current AI that controls systems like ChatGPT and other budding research assistants.
LLMs were the topic of the discussion panel, one of the final events of the day, which featured Gina Grossi, a Niagara College professor of computer programming in the Game Design program and current Ph.D. student in the Intelligent Systems and Data Science program at Brock; the aforementioned Jad Kabarra; Rahul Kammar, an assistant professor in the Brock Faculty of Education; and Ali Emami, an assistant professor in Brock’s Department of Computer Science.
The conversation was moderated by Robson De Grande, an associate professor in Brock’s Department of Computer Science.
Prompted by questions from both the audience and Dr. De Grande, the panellists discussed a myriad of topics about AI and LLMs, but the conversation mainly revolved around the strengths and weaknesses of LLMs, the use and ethicality of AI in education and what it truly meant to create an artificial intelligence.
It was made clear that LLMs like ChatGPT are not always factually accurate. They are merely programmed to create plausible conversation that could be coming from a human.
A system like Google is made for retrieval, meaning if a user searches for books on journalism, for example, the search results will be factually accurate. An LLM, on the other hand, does not have the same requirements. The same search could call up a list of books that all seem real but may not actually be so. As long as it sounds plausible, the system has done what it was programmed to do.
This makes its use in education rather tricky. Many Canadian school boards have unclear policies when it comes to the use of AI in the classroom, and in the US, some are integrating it into curriculums and others are outright banning it. The same arguments can be seen at Brock and other universities – some professors encourage its use while some consider any usage academic misconduct.
The predominant opinion coming from the panel was that AI isn’t going anywhere and banning it will not prepare students for their futures, but rather the ‘pasts of their teachers.’ Examples were given of pre-existing evolutions of education, such as when teachers had to adapt to the creation of the calculator in the form of word problems.
Importantly though, AI cannot replace education. The act of learning is meant to be a struggle: one that creates enduring intelligence in students that allows them to thrive in the world. Increased access to information does not render education moot, the panel agreed, for this was not the case when the Internet was invented.
The panel discussed how when tools like the calculator and the Internet were invented, the standard for education needed to be raised; now, it needs to be raised again to accommodate AI. As technology advances, so does intelligence. Humans learn when more elaborate tools are created, like how modern programmers have become more efficient than the original coders, who needed to input ones and zeroes rather than strings of text and commands.
This begs an important question, however, which the panel discussed at length: what is the standard for artificial intelligence? They described how the thresholds for programs to be declared intelligent are constantly being surpassed. In the past, if a program could play a game of chess it was considered intelligent. Recently, it’s been about emulating human language and now even that has been accomplished.
A recurring topic throughout the day was whether bias belongs in AI. Bias could be problematic for a system meant to deliver facts, but as the panel discussed, everybody has a bias, no matter how objective they try to be. So then, what is true intelligence? And as AI becomes more advanced, what becomes artificial about it?
Does the artificiality of AI mean it could be more intelligent than people? If yes, would humans even be comfortable creating such a thing if it meant surpassing the limits of humanity? The panellists could not definitively answer these questions, but they said that if they could, there would be no need for all their combined, continuous research.
AI Day ended with a speech from Donald Ziraldo, a leading figure in the Canadian wine-making industry. He discussed how large businesses are paying close attention to AI and implementing it where they can. It can make mundane tasks easier and accelerate productivity, and it even has uses in gene sequencing and combatting climate change, to name a few examples.
Dr. Lynn Wells, Provost and Vice-President Academic at Brock, gave the official closing remarks. After her speech, Dr. Betty Ombuki-Berman, a professor at the Brock Department of Computer Science, shared her desire for AI Day to be an ongoing event—especially if the proposed Centre of Artificial Intelligence comes to fruition.
While this event covered a lot about AI and LLMs, there was simply not enough time to cover everything, especially with the technology progressing at the rapid pace that it is. ChatGPT took the world by storm this past year. Should this event return next year, there will be plenty more to discuss.